Q-learning and SARSA algorithms

In this section we descuss off-line and on-line algorithms to compute
the optimal policy in case the exact model is not known.